Phan, Thien Vy (Scientific instructor)Nguyen, Truong Khanh Vy2023-03-232024-08-132023-03-232024-08-132020https://repository.ntt.edu.vn/handle/298300331/266979 p.Result and discussion: This study had totally five QSAR models with dependent value was MPC8, which were created by different descriptors. The validation results revealed that R2 values and Q2 value of all models were surprisingly greater than 0.80 and 0.77, respectively. In addition, the other values such as RMSE, R2pred, 𝑟𝑚 ̅̅2̅ and ∆r2m were come up to validation criteria. Two best models of five were LVFX and AZT_1 would be applied to virtual screening database. As a result, there were 15 compounds and 07 compounds had predicted MPC8 less than 10 µM in LVFX model and AZT_1 model, respectively, also, all of them met Five rules of Lipinski. On the other hand, five substances from Drugbank database were forecasted to have predicted MPC8 under 10 µM . Additionally, the predicted MPC8 value under 10 µM (G-9 (spergulin A) from Glinus oppositifolius.en2D-QSARMexAB-OprMPseudomonas aeruginosEfflux pump inhibitorsVirtual screeningMô hình 2D - QSARThesis